photonic integration on the hybrid silicon evanescent ... · on hybrid silicon evanescent devices....

18
Hindawi Publishing Corporation Advances in Optical Technologies Volume 2008, Article ID 682978, 17 pages doi:10.1155/2008/682978 Review Article Photonic Integration on the Hybrid Silicon Evanescent Device Platform Hyundai Park, 1 Alexander W. Fang, 1 Di Liang, 1 Ying-Hao Kuo, 1 Hsu-Hao Chang, 1 Brian R. Koch, 1 Hui-Wen Chen, 1 Matthew N. Sysak, 2 Richard Jones, 2 and John E. Bowers 1 1 Department of Electrical and Computer Engineering, University of California, Santa Barbara, CA 93106, USA 2 Intel Corporation, 2200 Mission College Blvd, SC12-326, Santa Clara, CA 95054, USA Correspondence should be addressed to John E. Bowers, [email protected] Received 17 December 2007; Revised 25 January 2008; Accepted 12 March 2008 Recommended by D. Lockwood This paper reviews the recent progress of hybrid silicon evanescent devices. The hybrid silicon evanescent device structure consists of III-V epitaxial layers transferred to silicon waveguides through a low-temperature wafer bonding process to achieve optical gain, absorption, and modulation eciently on a silicon photonics platform. The low-temperature wafer bonding process enables fusion of two dierent material systems without degradation of material quality and is scalable to wafer-level bonding. Lasers, amplifiers, photodetectors, and modulators have been demonstrated with this hybrid structure and integration of these individual components for improved optical functionality is also presented. This approach provides a unique way to build photonic active devices on silicon and should allow application of silicon photonic integrated circuits to optical telecommunication and optical interconnects. Copyright © 2008 Hyundai Park et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 1. INTRODUCTION Recent research in silicon photonics has been driven by the motivation to realize silicon optoelectronic integrated devices using large scale, low-cost, and highly accurate CMOS technology. Silicon is transparent at the 1.5 μm and 1.3 μm telecommunication wavelengths and has demon- strated low loss waveguide with losses in the range of 0.2 dB/cm 1 dB/cm. The large index contrast of silicon waveguides with silicon dioxide cladding results in highly confined optical modes and reduction of waveguide bend radii leading to dense photonic integration. This has resulted in advances in passive devices such as compact filters [1], optical buers [2], photonic crystals [3], and wavelength multiplexer/demultiplexers [4, 5]. It was only recently that silicon has been demonstrated as a high-speed modulator. Silicon-based modulators have been reported using free carrier plasma dispersion in Mach- Zehnder interferometer structure [6, 7], photonic crystals [8], and ring resonator structures [9]. Strained silicon has been shown to break the inversion symmetry of silicon allowing silicon to exhibit linear electro-optic refractive index modulation [10]. Recently, an electroabsorption mod- ulator on silicon has been demonstrated based on the quantum confined stark eect in strained silicon germanium [11]. Light detection is another major research topic in silicon photonics. Strained germanium and silicon germanium push the absorption out to 1.55 μm wavelength and are attractive since it is compatible with CMOS processing capabilities [12, 13]. Integration of the photodetector with the receiver is critical for lower capacitance and higher sensitivity [14]. The indirect bandgap of silicon has been a key hurdle for achieving optical gain elements. Although Raman lasers and amplifiers [1517], and optical gain in nanopatterned silicon have been observed [18], an electrically pumped silicon waveguide gain element has been an unsolved challenge. An alternative to fabricating the gain element in silicon is to take prefabricated lasers and couple them to silicon waveguides. However, due to the tight alignment tolerances of the optical modes and the need to align each laser individually, this method has limited scalability, and it is dicult to envision die attaching more than a few lasers to each chip without prohibitive expense. Furthermore, the reflections at the chip interfaces limit the gain and spectral flatness that can be achieved.

Upload: others

Post on 08-Jun-2021

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hindawi Publishing CorporationAdvances in Optical TechnologiesVolume 2008, Article ID 682978, 17 pagesdoi:10.1155/2008/682978

Review ArticlePhotonic Integration on the Hybrid SiliconEvanescent Device Platform

Hyundai Park,1 Alexander W. Fang,1 Di Liang,1 Ying-Hao Kuo,1 Hsu-Hao Chang,1 Brian R. Koch,1

Hui-Wen Chen,1 Matthew N. Sysak,2 Richard Jones,2 and John E. Bowers1

1 Department of Electrical and Computer Engineering, University of California, Santa Barbara, CA 93106, USA2 Intel Corporation, 2200 Mission College Blvd, SC12-326, Santa Clara, CA 95054, USA

Correspondence should be addressed to John E. Bowers, [email protected]

Received 17 December 2007; Revised 25 January 2008; Accepted 12 March 2008

Recommended by D. Lockwood

This paper reviews the recent progress of hybrid silicon evanescent devices. The hybrid silicon evanescent device structure consistsof III-V epitaxial layers transferred to silicon waveguides through a low-temperature wafer bonding process to achieve opticalgain, absorption, and modulation efficiently on a silicon photonics platform. The low-temperature wafer bonding process enablesfusion of two different material systems without degradation of material quality and is scalable to wafer-level bonding. Lasers,amplifiers, photodetectors, and modulators have been demonstrated with this hybrid structure and integration of these individualcomponents for improved optical functionality is also presented. This approach provides a unique way to build photonic activedevices on silicon and should allow application of silicon photonic integrated circuits to optical telecommunication and opticalinterconnects.

Copyright © 2008 Hyundai Park et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

1. INTRODUCTION

Recent research in silicon photonics has been driven bythe motivation to realize silicon optoelectronic integrateddevices using large scale, low-cost, and highly accurateCMOS technology. Silicon is transparent at the 1.5 μm and1.3 μm telecommunication wavelengths and has demon-strated low loss waveguide with losses in the range of0.2 dB/cm ∼ 1 dB/cm. The large index contrast of siliconwaveguides with silicon dioxide cladding results in highlyconfined optical modes and reduction of waveguide bendradii leading to dense photonic integration. This has resultedin advances in passive devices such as compact filters [1],optical buffers [2], photonic crystals [3], and wavelengthmultiplexer/demultiplexers [4, 5].

It was only recently that silicon has been demonstratedas a high-speed modulator. Silicon-based modulators havebeen reported using free carrier plasma dispersion in Mach-Zehnder interferometer structure [6, 7], photonic crystals[8], and ring resonator structures [9]. Strained silicon hasbeen shown to break the inversion symmetry of siliconallowing silicon to exhibit linear electro-optic refractiveindex modulation [10]. Recently, an electroabsorption mod-

ulator on silicon has been demonstrated based on thequantum confined stark effect in strained silicon germanium[11].

Light detection is another major research topic in siliconphotonics. Strained germanium and silicon germanium pushthe absorption out to 1.55 μm wavelength and are attractivesince it is compatible with CMOS processing capabilities[12, 13]. Integration of the photodetector with the receiveris critical for lower capacitance and higher sensitivity [14].

The indirect bandgap of silicon has been a key hurdle forachieving optical gain elements. Although Raman lasers andamplifiers [15–17], and optical gain in nanopatterned siliconhave been observed [18], an electrically pumped siliconwaveguide gain element has been an unsolved challenge.

An alternative to fabricating the gain element in siliconis to take prefabricated lasers and couple them to siliconwaveguides. However, due to the tight alignment tolerancesof the optical modes and the need to align each laserindividually, this method has limited scalability, and it isdifficult to envision die attaching more than a few lasersto each chip without prohibitive expense. Furthermore, thereflections at the chip interfaces limit the gain and spectralflatness that can be achieved.

Page 2: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

2 Advances in Optical Technologies

P-InGaAsP-InP cladding

P-AlGaInAs SCHAlGaInAs MQWs

n-InPn-InP/InGaAsP

N pad

P pad

Insulation layer

H+ H+

Air gapBuried oxide

Si substrate

III-Vregion

SOIregion

(a)

N pad

P pad

Insulation layer

Buried oxide

Si substrate

III-Vregion

SOIregion

(b)

Figure 1: Hybrid silicon evanescent device cross-section structures: (a) a wide III-V mesa for amplifiers and lasers, (b) a narrow III-V mesafor detectors and modulators.

Epitaxial growth of III-V layers on silicon substrateshas been investigated as well. The demonstrations includeInGaAs quantum dot lasers [18] and InGaSb quantum welllasers [19] fabricated on silicon substrate. Those demonstra-tions have widened the possibility of building monolithicallyintegrated on-chip laser sources on the silicon photonicsplatform. However, an efficient coupling scheme from theIII-V lasers to the silicon waveguide needs to be developedsince most of the lasing mode is in the III-V layers.

Recently, we have demonstrated a hybrid integrationplatform utilizing III-V epitaxial layers transferred to siliconto realize many types of photonic active devices through asingle wafer bonding step. The wafer-bonded structure formsa hybrid waveguide, where its optical mode lies both in sili-con and III-V layers. This structure enables the use of III-Vlayers for active light manipulation such as gain, absorption,and electro-optical effect for the amplifiers, lasers, detectors,and modulators. In this paper, we review the recent progresson hybrid silicon evanescent devices. In Section 2, the devicestructure and design issues are introduced. In Section 3, thedevice fabrication process is described. Section 4 presentsthe performance and characteristics of fabricated siliconevanescent devices. Finally, several potential future paths ofthis research are discussed in Section 5.

2. DEVICE PLATFORM

Figure 1 shows the general structure of hybrid siliconevanescent devices. The hybrid structure is comprised ofa III-V region bonded to a silicon waveguide fabricatedon a silicon-on-insulator wafer. The mesa structure formedon the III-V region enables the current flow through themultiple quantum well region. The general structure ofIII-V layers consists of a p-type contact layer, a p-typecladding, a p-type separated confinement heterostructure(SCH) layer, an undoped multiple quantum well layer, n-type contact layer, and n-type super lattices. Amplifiers andlasers have a wide III-V mesa (12 μm ∼ 14 μm) for betterheat conduction and mechanical strength (Figure 1(a)) whilea narrow III-V mesa (2 μm ∼ 4 μm) is chosen for detectorsand modulators for high-speed operation with a reducedcapacitance (Figure 1(b)). The optical mode in this hybridwaveguide lies both in the silicon waveguide and the multiplequantum well layers. The confinement factors in III-V andsilicon regions of the hybrid waveguide can be manipulated

by changing the silicon waveguide dimensions. The quantumwell confinement factor is a critical design parameter inorder to achieve enough optical gain and absorption whilethe silicon confinement factor is an important parameterdetermining coupling efficiency when the device is integratedwith silicon passive devices. Figure 2 shows three differentmode profiles with three different waveguide widths. Ingeneral, the silicon confinement factor increases as theheight or width of the silicon waveguide increases while thequantum well confinement factor decreases. The epitaxialstructures and confinement factors for each device set willbe specified in Section 4.

3. FABRICATION

3.1. Plasma assisted low-temperature wafer bonding

The transfer of the indium phosphide (InP)-based epitaxiallayer structure to the silicon-on-insulator (SOI) substrateis a key step in the fabrication of this hybrid platformand has direct impact on the device performance, yield,and reliability. Due to the mismatch between the thermalexpansion coefficient of silicon and indium phosphide (αSi

= 2.6 ×10−6/K , αInP = 4.8 ×10−6/K), high-temperature(>400◦C) annealing steps are not desirable. Figure 3(a)shows a Nomarski photograph of the top surface of an InPdie transferred to a silicon-on-insulator substrate at 600◦C.Crosshatching can be seen for this high-temperature directwafer bonding which can lead to degradation of materialquality and scalability issues due to the accumulation ofstress over larger sample sizes. In order to resolve thisissue, low-temperature annealing is used with an oxygenplasma surface treatment to enable strong bonding [20].An annealing temperature of 300◦C is chosen to minimizebonding stress while still being able to convert the weakHydrogen bonds formed by the room temperature bondingto strong covalent Si–O–In and Si–O–P bonds. Figure 3(b)shows a successful transfer of InP epitaxial layers to SOI withsmooth device quality surface morphology and no interfacialvoids.

Figure 4 is the schematic process flow of the oxygenplasma assisted low-temperature (300◦C) wafer bonding.After rigorous sample cleaning and close microscopic inspec-tion with 200x magnification, the native oxide on SOI andInP are removed in standard buffer HF solution (1HF : 7

Page 3: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 3

43.5

32.5

21.5

10.5

0−0.5

y(μ

m)

−3 −2 −1 0 1 2 3

x (μm)

W = 1 μm

(a)

43.5

32.5

21.5

10.5

0−0.5

y(μ

m)

−3 −2 −1 0 1 2 3

x (μm)

W = 1.25 μm

(b)

43.5

32.5

21.5

10.5

0−0.5

y(μ

m)

−3 −2 −1 0 1 2 3

x (μm)

W = 2 μm

(c)

Figure 2: Mode profiles with different waveguide widths. Height of the silicon waveguide is fixed at 0.7 μm.

(a) (b)

Figure 3: Nomarski microscope images of the transferred III-Vsurface at bonding temperatures of (a) 600◦C and (b) 300◦C.

H2O) and NH4OH (39%), respectively, resulting in clean,hydrophobic surfaces. The samples then undergo an oxygenplasma surface treatment to grow an ultrathin layer of oxide(<5 nm) [21] which leads to very smooth (rms roughness<0.5 nm) hydrophilic surfaces, which is less sensitive tothe microroughness as compared to hydrophobic bonding.The Si–O–Si bonds of the oxide (SOI side) are also foundto be more strained than conventional oxides formed instandard RCA-1 cleaning process or other hydrophilic wet-chemical treatment, and have a higher readiness to breakand form new bonds. O2 energetic ion bombardment alsoacts as a final cleaning step to remove hydrocarbons andwater-related species on the sample surface efficiently. Thefollowing deionized water dip further terminates the oxidesurface by polar hydroxyl groups OH−, forming bridgingbonds between the mating surfaces to result in spontaneousbonding at room temperature [22]. To strengthen the bond,the bonded sample is placed in a conventional wafer bondingmachine (Suss Bonder SB6E), where the samples are heldtogether at a pressure of 1.5 MPa and a temperature of 300◦C,under vacuum (<4 × 10−4 Torr) from 1 to 12 hours. The300◦C annealing process enhances out diffusion of moleculestrapped at the interface and desorption of chemisorbed sur-face atoms, such as hydrogen, while activating the formationof covalent bonds to achieve higher bonding energy [20].After annealing and cooling, the InP substrate is selectivelyremoved in a 3HCl : 1H2O solution at room temperature.

Figure 5 shows a 2-inch InP-based expitaxial waferbonded on a SOI sample cleaved from a 6 inch SOIwafer. Smooth III-V morphology with no interfacial void is

achieved in the bonded area. The two defects on the left-handside of the figure are due to wafer handling with tweezersand InP epitaxial surface defects that are 29 μm in diameter.Successful epitaxial transfer on 2 inch wafer demonstrates thescalability of this oxygen plasma-assisted low-temperaturebonding process, which subsequently paves the way for massproduction of the hybrid devices.

3.2. Silicon waveguide and III-V back-end processing

The general procedure of silicon waveguide formation on anSOI wafer and III-V back-end processing after wafer bondingprocess is as follows. The silicon waveguide is formedon the (100) surface of an undoped silicon-on-insulator(SOI) substrate using Cl2/Ar/HBr-based plasma reactive ionetching. The thickness of the buried oxide (BOX) is 1 μm forthe devices reported in this paper. The III-V epitaxial layeris then transferred to the patterned silicon wafer throughlow-temperature oxygen plasma-assisted wafer bonding,which was described in Section 3.1. After removal of theInP substrate, mesa structures on III-V layers are formedby dry-etching the p-type layers using a CH4/H/Ar-basedplasma reactive ion etch. Subsequent wet-etching of thequantum well layers to the n-type layers is performed usingH3PO4/H2O2. Ni/AuGe/Ni/Au alloy contacts are depositedonto the exposed n-type InP layer. Pd/Ti/Pd/Au p-contactsare then deposited on the center of the mesas. For lasers andamplifiers, protons (H+) are implanted on the two sides ofthe p-type mesa to create a 4 μm wide current channel andto prevent lateral current spreading, ensuring a large overlapbetween the carriers and the optical mode. Ti/Au probe padsare then deposited on the top of the mesa. Then, if necessary,the sample is diced into bars and each bar is polished.

4. DEVICE RESULTS

4.1. Silicon evanescent lasers

4.1.1. 1550 nm Fabry Perot lasers

The first demonstrated device using the silicon evanescentdevice platform was the Fabry-Perot (FP) hybrid siliconevanescent laser [23]. The cavity for these lasers was madeby dicing the ends of the hybrid waveguide and polishing

Page 4: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

4 Advances in Optical Technologies

CleanIII-V

Device epiInGaAsInP sub.

SOI

O2 plasmasurface

treatment

O2 plasma

H2O dip &spontaneous

wafer bonding

Anneal

InP substrate selectiveremoval

in wet etch

Si Si

O

10 nm thick highlystrained SiO2

300◦C, 1–2 hrs4× 10−4 Torr

1.5 MPa

Post devicefabrication

Figure 4: Oxygen plasma-assisted low-temperature wafer bonding process flow.

Figure 5: Photograph of a 2 inch III-V epitaxial transferredto an SOI sample after 300◦C anneal for 2 hours and selectiveremoval of the InP substrate. InP wafer sidewall is surrounded byCrystalbondTM wax to protect InP device layer from being etchedlaterally during the substrate removal step. The white region in themiddle of the wafer is a reflection of the illuminator.

them to a mirror finish. The device presented here hastwo major changes from the first reported devices. First,the buried oxide thickness is reduced to 1 μm in order toreduce the thermal impedance of the device. Second, the III-V mesa was reduced to 12 μm in order to reduce the deviceseries resistance as shown in Figure 1(a). The waveguideheight, width, rib etch depth, and cavity length were 0.7 μm,

2 μm, 0.5 μm, and 850 μm, respectively. The calculatedconfinement factors in the silicon and the quantum wellregion are 63% and 4%, respectively. The epitaxial structureof the lasers is specified in Table 1.

The continuous wave (CW) LI curve for this device iscollected on one side with an integrating sphere as shownin Figure 6(a). In order to account for light exiting bothsides of the cavity, the data is multiplied by two. It canbe seen that the maximum laser output power, threshold,and differential efficiency at 15◦C are 24 mW, 70 mA, and16%, respectively. The device shows improvement in outputpower and differential efficiency while maintaining a similarthreshold when compared to the first generation device. Themaximum operating temperature is 45◦C.

Figure 6(b) shows a set of pulsed single-sided outputpower as a function of applied current (1 kHz repetition rate,0.1% duty cycle) for stage temperatures ranging from 15to 50◦C. The characteristic temperature (T0) and an abovethreshold characteristic temperature (T1) [24] are 60 K and120 K, respectively. The thermal impedance of the laser ismeasured using a combination of two experiments. The firstset of measurements is used to establish a baseline for theshift in lasing wavelength as a function of active region(stage) temperature (dλ/dT) as shown in Figure 7(a). Similarto the characteristic temperature measurement above, thisexperiment is performed pulsed to ensure that there isminimal device heating other than what is provided by

Page 5: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 5

25

20

15

10

5

0

Dou

ble

side

dto

talo

utp

ut

pow

er(m

W)

0 100 200 300 400 500

Current (mA)

15◦C20◦C25◦C30◦C

35◦C40◦C45◦C

(a)

20

15

10

5

0

Dou

ble

side

dpu

lsed

outp

ut

pow

er(m

W)

0 20 40 60 80 100 120 140 160 180 200

Current (mA)

15◦C20◦C25◦C30◦C

35◦C40◦C45◦C50◦C

(b)

Figure 6: LI curves for a 1550 nm FP silicon evanescent laser (a) double-sided continuous wave output power, (b) double-sided pulsedoutput power.

Table 1: III-V epitaxial layer structure with a 1550 nm photoluminescence peak.

Name Composition Doping concentration Thickness

P contact layer P-type In0.53Ga0.47As 1× 1019 cm−3 0.1 μm

Cladding P-type InP 1× 1018 cm−3 1.5 μm

SCH P-type Al0.131Ga0.34In0.528 As, 1.3Q 1× 1017 cm−3 0.25 μm

Quantum wellsAl0.089Ga0.461In0.45As, 1.3Q(9x) undoped 10 nm

Al0.055Ga0.292In0.653As, 1.7Q(8x) undoped 7 nm

N layer N-type InP 1× 1018 cm−3 110 nm

Super latticeN-type In0.85Ga0.15As0.327P0.673, 1.1Q(2x) 1× 1018 cm−3 7.5 nm

N-type InP (2x) 1× 1018 cm−3 7.5 nm

N bonding layer N-type InP 1× 1018 cm−3 10 nm

the temperature-controlled stage. The second measurementis performed CW, and is used to measure the shift inwavelength as a function of applied electrical power tothe laser (dλ/dP) as shown in Figure 7(b). The thermalimpedance ZT is then given by (1)

ZT =(dλ

dT

)−1( dλdP

). (1)

In both the dλ/dT and dλ/dP experiments, a singlelongitudinal mode in the laser spectrum is monitored.Combining the results from Figures 7(a) and 7(b), the laserthermal impedance is measured to be 42◦C/W.

The thermal performance of the hybrid laser depends onseveral factors. These include the amount and location ofheat that is generated, the thermal conductivity of the layerssurrounding the heat sources, and the operating temperatureof the laser active region. To model the temperature riseas a function of applied bias, we have employed a two-

dimensional finite element modeling technique. For thehybrid laser cross-section shown in Figure 1(a), there aresix major sources of thermal energy. These include resistiveheating in the p-cladding, the n-contact layer, the activeregion, and the p and n contacts, along with heat generatedby the diode drop associated with the active region. Moredetailed information about the values used in the simulationcan be found in [25].

A two dimensional temperature profile of the hybrid laseroperating at 500 mA is shown in Figure 8(a). The dissipatedelectrical power and the predicted temperature rise in thelaser active region are plotted as a function of applied currentin Figure 8(b). The contribution of each layer to the totalelectrical power dissipation is also shown in the same figure.Combining the simulated temperature rise, the dissipatedelectrical power, and the output optical power, the thermalimpedance of the laser is 43.5◦C/W, which is within 5% ofour initial experimental results.

Page 6: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

6 Advances in Optical Technologies

1569.2

1569

1568.8

1568.6

1568.4

1568.2

1568

FPm

ode

wav

elen

gth

(nm

)

0 10 20 30

Stage temperature T (◦C)

Δλ/ΔT = 0.067 nm/ ◦C

Δλ−40

−50

−60

−70

−80

15◦C 30◦CIn

ten

sity

(arb

.)

Wavelength (nm)1564 1569 1574

(a)

1569.5

1569.3

1569.1

1568.9

1568.7

FPm

ode

wav

elen

gth

(nm

)

0 0.1 0.2 0.3 0.4

Dissipated power (W)

Δλ/ΔP = 2.8 nm/W

Δλ

0

−20

−40

−60

−80

165 mW

232 mW

Inte

nsi

ty(a

rb.)

Wavelength (nm)1568 1569 1569 1570 1570

(b)

Figure 7: (a) Pulsed measurement results for the shift in lasing wavelength (single FP mode) as a function of stage temperature. The insetcontains a laser output spectrum at 15◦C and 30◦C. (b) Results for the shift in lasing wavelength (single FP mode) as a function of dissipatedpower.

5

4

3

2

1

0

−1

−2

−3

−4

−5

Posi

tion

(μm

)

−10 −5 0 5 10

Position (μm)

Si

SiO2

Si

Active region

Au metal

p-InP

n-InPInP

n-InPInP

Max: 67.2

60

50

40

30

20

10

0Min: 0

Surface: temperature (T)

(a)

2500

2000

1500

1000

500

0

Dis

sipa

ted

pow

er(m

W)

0 100 200 300 400 500 600

Input current (mA)

ΔTmax(1μm oxide)

ΔTmax(no oxide)

Active region power dissipation

p-InP power dissipation

n-InP power dissipation

Total power dissipation

70

60

50

40

30

20

10

0

Tem

per

atu

re(◦

C)

(b)

Figure 8: (a) Two-dimensional temperature profile in the hybrid laser at a bias current of 500 mA. (b) Theoretical predictions for thedissipated electrical power in the various laser sections along with the predicted temperature rise in the active region as a function of contactcurrent.

To show the effect of the buried oxide on the thermalimpedance, Figure 8(b) also includes a simulation of thetemperature rise in the device when the buried oxide hasbeen removed. Reducing the thickness of the BOX layerresults in lowering the thermal impedance to 18◦C/W. Thisillustrates how the high thermal conductivity of siliconshould result in the very low thermal resistances of siliconphotonic devices.

4.1.2. 1310 nm Fabry Perot lasers

1310 nm hybrid silicon lasers are also important for manydata and telecommunication applications [26]. The epitaxiallayer structure used here (Table 2) contains an electron

blocking layer between the quantum wells and the SCH layerfor better injection efficiency [27]. The layer is designed toprovide a high conduction band offset between the barrierand the SCH layer to prevent electrons from leaking out ofthe quantum well region while a valence band offset is keptlow not to alter the hole flow into the quantum wells.

Lasers with fundamental transverse mode or withsecond-order transverse mode can be designed and fab-ricated. Figure 9(a) shows the simulated quantum wellconfinement factor as a function of silicon waveguide heightwhile keeping the waveguide width and slab thickness at2.5 μm and 0.2 μm, respectively. Quantum well confinementrepresents the mode overlap to the 4 μm wide quantum wellat the center, where electrons and holes are injected. In the

Page 7: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 7

Table 2: III-V epitaxial layer structure with a 1303 nm photoluminescence peak.

Name Composition Doping Concentration Thickness

P contact layer P-type In0.53Ga0.47As 1× 1019 cm−3 0.1 μm

Cladding P-type InP 1× 1018 cm−3 1.5 μm

SCH P-type Al0.4055Ga0.064In0.5305 As, 0.9Q 1× 1017 cm−3 0.25 μm

Electron blocking layer Al0.4764Ga0.0189In0.5047As undoped 10 nm

Quantum wellsAl0.055Ga0.292In0.653As, 1.0Q(9x) undoped 10 nm

Al0.178Ga0.1234In0.0986As, 1.5Q(8x) undoped 7 nm

N layer N-type InP 1× 1018 cm−3 110 nm

Super latticeN-type In0.85Ga0.15As0.327P0.673, 1.1Q(2x) 1× 1018 cm−3 7.5 nm

N-type InP (2x) 1× 1018 cm−3 7.5 nm

N bonding layer N-type InP 1× 1018 cm−3 10 nm

0.1

0.05

0Qu

antu

mw

ellc

onfi

nem

ent

fact

or

0.4 0.45 0.5 0.55 0.6 0.65

Silicon height (μm)

Fundamental modeSecond transverse mode

(a)

Fundamental Second transverse

H = 0.7μm

H = 0.475μm

H = 0.4μm

(b)

Figure 9: The confinement factor in III-V region simulationfor different silicon waveguide height (a) confinement factorcalculations for the fundamental mode and second transverse modeas a function of waveguide height (b) fundamental and secondtransverse modes for waveguide heights of 0.7, 0.475, and 0.4 μm.The image aspect ratio (Height:Width) is 4 : 1.

tall silicon height regime (right hand part in Figure 9(a)),a higher quantum well confinement factor for the secondtransverse mode exists. In the short silicon height regime(left-hand part in Figure 9(a)), the fundamental mode has a

(a) (b)

(c) (d)

Figure 10: The simulated and measured lasing optical modeprofiles (a) simulated optical mode for a waveguide height of0.4 μm, (b) simulated optical mode for a waveguide height of0.7 μm, (c) measured optical mode for a waveguide height of0.4 μm, and (d) measured optical mode for a waveguide height of0.7 μm.

higher quantum well confinement factor. Figure 9(b) showsthe fundamental mode and second transverse mode withdifferent waveguide heights illustrating the fundamentalmode increasingly lies more in the III-V region as the siliconheight decreases. The second transverse mode also undergoesan increase in III-V confinement factor but splits into twolateral lobes at lower silicon heights. This splitting reducesthe modal overlap with the 4 μm wide excited quantum wellregion at the center.

Two different silicon waveguide heights of 0.4 μm and0.7 μm have been chosen to study the lasing mode selectiondepending on different quantum well confinement factors.The width and slab height of the silicon waveguide is 2.5 μmand 0.2 μm, respectively. The device length is ∼850 μm. Fig-ures 10(a) and 10(b) show the simulated mode profiles withthe largest quantum well confinement factor for waveguideheights of 0.4 μm and 0.7 μm, respectively. The measuredlasing mode profiles in Figures 10(c) and 10(d) agree with thesimulation results, indicating the quantum well confinement

Page 8: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

8 Advances in Optical Technologies

8

6

4

2

0

Sin

gle

side

dfi

ber

cou

pled

outp

ut

pow

er(m

W)

0 50 100 200 300 400 500

Current (mA)

15◦C25◦C35◦C

45◦C65◦C

Dou

ble

side

dto

talo

utp

ut

pow

er(m

W)

60.4

37.8

25.2

12.6

0

(a)

6

4

2

0

Sin

gle

side

dfi

ber

cou

pled

outp

ut

pow

er(m

W)

0 20 40 60 80 100 120

Current (mA)

15◦C25◦C75◦C

85◦C95◦C105◦C

Dou

ble

side

dto

talo

utp

ut

pow

er(m

W)

37.8

25.2

12.6

0

(b)

Figure 11: LI curves for 1310 nm FP silicon evanescent lasers (a) fundamental mode lasing device, (b) second transverse mode lasing device.The secondary y-axis represents the estimated total laser output power.

factor primarily determines the lasing mode and can beengineered by changing the silicon waveguide dimensions.

Figure 11 shows the measured single-sided fiber-coupledCW output power as a function of the injected current atdifferent temperatures for two different lasers (0.4 μm and0.7 μm waveguide heights). The threshold current for bothdevices at 15◦C is 30 mA. The output power at 100 mAis 2.9 mW and 5 mW and its corresponding differentialquantum efficiency is 2.5% and 8% for the fundamentalmode and second transverse mode lasing devices, respec-tively. The estimated double-sided total output power isshown in the secondary y-axis taking account for the outputpower from the both facets and a coupling loss of 5 dBbetween the device and the lensed fiber. The double sidedtotal output power at 100 mA is estimated to be 18 mWand 31 mW, and its total differential quantum efficiencyis 15% and 50% for the fundamental mode and secondtransverse mode lasing devices, respectively. The device las-ing with second transverse mode (0.7 μm waveguide height)operates up to 105◦C and has better performance than thedevice lasing with a fundamental mode primarily due tothe higher quantum well confinement factor (10% versus7%). Moreover, the overall performance of 1.3 μm lasersis superior to 1.5 μm lasers previously described becausethe carrier blocking layer is incorporated and becauseof the reduced intravalence band absorption and Augerscattering.

4.1.3. 1550 nm integrated racetrack laserand photodetectors

Figure 12(a) shows the layout of an integrated hybrid siliconevanescent racetrack laser and two photodetectors operatingat 1550 nm [28]. The same epitaxial structure described inTable 1 is used both for the laser and the detector. This laser

does not rely on facet dicing or polishing and can be testedon-chip with simple probing of the laser and photodetectors.

The waveguide height, width, and rib etch depth were0.69 μm, 1.5 μm, and 0.5 μm, respectively. The scanningelectron microscope (SEM) image of the fabricated devicesis shown in Figure 12(b). It consists of a racetrack ring res-onator with a straight waveguide length of 700 μm and ringradii of 200 and 100 μm. A directional coupler is formed onthe bottom arm by placing a bus waveguide 0.5 micron awayfrom the racetrack. Since clockwise and counterclockwisepropagating modes of ring lasers are only weakly coupled,two 440 μm long photodetectors are used to collect the laserpower; the clockwise being collected at the left detector, andthe counterclockwise being collected at the right detector.These photodetectors have the same waveguide structure asthe hybrid laser, and the only difference being that they arereverse biased to collect photogenerated carriers.

To estimate the laser output power from the measuredphotocurrent, the responsivity of the detector is first mea-sured by dicing and polishing a discrete detector in the samechip and launching laser light into the detector through alensed fiber. The fiber coupled responsivity was measuredto be 0.25 A/W at 1580 nm. Taking into consideration the∼30% reflection off the waveguide facet and an estimated5.25 +/− .25 dB coupling loss, the photodetector responsivityis estimated to be in the range of 1.25–1.11 A/W. This corre-sponds to an internal quantum efficiency of around 92%. Aresponsivity of 1.25 A/W is used for laser power estimationsuch that the laser power values are on the conservativeside. These measurement results are also consistent with themeasured responsivity from stand alone photodetectors [29].

The total laser output power collected at both detectorsas a function of current and temperature is shown inFigure 12(c) for a laser with a ring radius of 200 μm and acoupler interaction length of 400 μm. The laser has a total

Page 9: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 9

P-metal N-metal

Laser

Coupling length

Directional couplerPhotodetector

PD1Photodetector

PD2

(a)

500 μm

(b)

30

25

20

15

10

5

0

Dou

ble

side

dto

talo

utp

ut

pow

er(m

W)

0 100 200 300 400 500

Current (mA)

15◦C25◦C35◦C45◦C

55◦C60◦C66◦C

(c)

−75

−80

−85

−90

−95

−100

−105

A.u

.(dB

)

1590 1591 1592 1593 1594 1595

Wavelength (nm)

(d)

Figure 12: (a) The layout of the racetrack resonator and the photodetectors. (b) A top view SEM micrograph of two racetrack resonatorlasers. The racetrack resonator lasers on the top and bottom have radii of 200 and 100 μm, respectively. (c) The LI curve for a laser with abend radius, R, of 200 μm, and directional coupler coupling length, Linteraction, of 400 μm for various temperatures. (d) The spectrum for alaser with R = 100 μm and Linteraction = 400 μm.

output power of 29 mW with a maximum lasing temperatureof 60◦C. The differential efficiency is 17% and the laserthreshold is 175 mA at 15◦C. The laser spectrum is shown inFigure 12(d) with its lasing peak in the range of 1592.5 nm.

Since the two modes of propagation are not coupledand are degenerate, mode competition is typical in suchlasers with a ring cavity structure. Figure 13(a) shows thephotocurrent measured separately from the left (PD1) andthe right (PD2) detectors as a function of the laser drivecurrent for a laser with a ring radius of 100 μm and a couplerinteraction length of 100 μm. There are two distinct regionsof operation above threshold: (1) unidirectional bistable(low-bias current), and (2) alternating oscillation (high-biascurrent). In the unidirectional bistable region, the laser islasing in either one direction or the other and the laser outputswitches from one direction to the other direction as the laserdiode current increase. At higher bias (>360 mA), the laserenter the alternating oscillation region, the output power

exhibit oscillatory behavior. The unidirectional bistability ofthe racetrack laser can be alleviated if one of the detectors isforward biased to inject light into the cavity (Figure 13(b)).The forward-biased photodetector acts as an ASE source,which increases the photon density of the clockwise mode,leading to greater simulated emission and mode selection.This shows that the lasing direction can be controlled byforward biasing one of the photodiodes; that is, lasing eitherclockwise or counterclockwise.

4.1.4. 1550 nm mode locked lasers

Silicon hybrid lasers can be mode locked at a variety offrequencies from 10 GHz to 40 GHz [30], for potentialapplications in optical pulse generation, OTDM, WDM, andregenerative all-optical clock recovery [31].

An FP ML-SEL is shown in Figure 14(a) and its test setupis shown in Figure 14(b). The 39.4 GHz FP ML-SEL had

Page 10: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

10 Advances in Optical Technologies

3

2.5

2

1.5

1

0.5

00 100 200 300 400 500

ILD (mA)

I PD

(mA

)

PD1PD2

(a)

3

2.5

2

1.5

1

0.5

00 100 200 300 400 500

ILD (mA)

I PD

(mA

)

IPD2 = 50 mA

(b)

Figure 13: (a) L-I curves of CW and CCW direction, (b) L-I curve of CW direction when the PD2 is forward bias with 50 mA.

a total cavity length of 1060 μm and a saturable absorber(SA) length of 70 μm. For all devices presented here, separategain and SA sections were electrically isolated using protonimplantation. Passive mode locking was achieved for a rangeof gain currents between 195 mA and 245 mA with similaroutput characteristics. For a gain current of 206 mA and asaturable absorber reverse bias voltage of 0.4 V, the pulsehas a sech2 shape with 4.2 picoseconds full width at halfmaximum (FWHM) pulsewidth, extinction ratio (ER) over18 dB between the peak and null, peak power of 4.5 dBmin fiber, and FWHM optical spectral width of 0.9 nm. Thetime bandwidth product is 0.4, close to the transform limitedvalue of 0.32 for sech2 pulses, indicating minimal chirp. Thelaser was capable of stable mode locking for a range of gaincurrents between 195 mA and 245 mA with similar outputcharacteristics.

By applying an RF signal to the saturable absorbersection, hybrid mode locking occurs and the jitter of thepulses can be considerably reduced [32] without changingthe pulsewidth or spectral width. To achieve subharmonichybrid mode locking, a 20 GHz RF source with 17 dBm of RFinput power is used. For these conditions, the absolute jitterof this laser is 1 picosecond and the locking range was 5 MHz.For all measurements in this paper, the jitter was evaluated byintegrating two times the single-sideband noise from 1 kHzto 100 MHz offset from the carrier frequency.

Increasing the cavity length to 4.16 mm resulted in10 GHz mode locking. The saturable absorber was 80 μmlong and the cavity was divided into 4 equal length gainsections that can be biased differently depending on theapplication. The bias conditions used for pulse generationwere as follows: Gain 1 and Gain 2 (adjacent to the saturableabsorber) were biased together at 531 mA, Gain 3 was biasedat 140 mA, Gain 4 at 149 mA, and the SA was biased with

Polished facet n metal bridge

Gain section n metal Optical waveguide

Absorber n contact Gain section n metal contact

Absorber p contact Gain section p metal

(a)

RF source

Mode locked siliconevanescent laser

EDFA

RFSA

OSA

SHG

PD

Splitter

Splitter

OpticalElectrical

(b)

Figure 14: (a) Scanning electron micrograph of one end of an FPML-SEL. (b) Schematic of the general experimental setup. For somemeasurements, more components were required as detailed in thereferences. EDFA = erbium doped fiber amplifier, SHG = secondharmonic generation autocorrelator, OSA = optical spectrumanalyzer, RFSA = 41 GHz radio frequency spectrum analyzer, PD= 40 GHz photodetector.

Page 11: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 11In

ten

sity

(dB

)

0

−10

−20

Pow

er(d

B)

0

−20

−40

−10 −5 0 5 10

Time (ps)

DataSech2 fit

1584 1586 1588 1590

Wavelength (nm)

(a)

10

9

8

7

6

5

4

3

2

1

0

Jitt

er(p

s)

−15 −10 −5 0 5 10 15 20

RF input power (dBm)

Passive (no RF)

Residual jitter,10 GHz RF drive

Absolute jitter,10 GHz RF drive

Absolute jitter,2.5 GHz RF drive

(b)

Figure 15: (a) Wavelength spectrum and autocorrelation trace ofthe 40 GHz FP ML-SEL output. The fit to the optical spectrumis the Fourier transform of the autocorrelation fit. (b) Jittermeasurements for the 10 GHz FP ML-SEL versus the RF inputpower.

−2.3 V. The FWHM pulsewidth was 3.9 picoseconds with asech2 shape, the ER is over 18 dB, the peak power in fiberwas 11.9 dBm, and the FWHM spectral width was 3.8 nm,containing 45 modes evenly spaced at 10.16 GHz. The timebandwidth product is 1.7, indicating significant chirping.Chirp can be reduced in future designs by incorporatingpassive silicon waveguide sections in the laser cavity with ashorter gain section [33]. This is fairly easy to do on thisplatform and is a potential advantage of the platform forlower repetition rate mode-locked lasers. Changing the SAbias between −0.5 and −2 V changes the output pulsewidthbetween 9 and 4 picoseconds, while the ER and output power

do not change significantly. Figure 15(b) shows results forpassive and hybrid mode locking with different RF powersand RF injection frequencies. Also shown is the residual jitter,which is the jitter of the hybrid mode locked laser comparedto that of the RF drive source. The minimum residual jitteris 199 femtoseconds with 15 dBm RF drive at 10.16 GHz. Atthis RF power, the locking range was 260 MHz.

Increasing the combined gain section bias currents to atotal of 1 A generates over 100 optical modes with powerlevels within 10 dB of the peak mode power. If the modeshave high enough quality, the single ML-SEL could becombined with an AWG and used as a multiple wavelengthsource for wavelength division multiplexing applications[34]. Across 100 modes, the linewidth was below 500 MHzand the OSNR was near 15 dB. The linewidths of thesemodes are too large for some applications, but for shortreach applications the signal quality is sufficient. The modequality can be improved by injecting a stable CW laserinto the hybrid mode-locked laser, inducing optical injectionlocking [34]. In this case, the spectral width narrows to30 modes within 10 dB, but the linewidth of each of thesemodes is reduced to that of the injected signal (<100 kHz).Also the OSNR is improved to over 25 dB for the majorityof the modes. It is expected that better stabilization of thehybrid mode-locked laser through packaging would allowfor a wider spectral width under optical injection locking,allowing for more high-quality modes to be generated.

A 30.4 GHz racetrack ML-SEL [31] is shown inFigure 16(a). A racetrack mode-locked laser has the advan-tage that its cavity is defined by lithography. Since the repeti-tion rate is determined by the cavity length, this allows for therepetition rate to be precisely determined and repeated acrossdifferent devices. Racetrack lasers can also be integratedmonolithically with other components. This laser has acavity length of 2.6 mm, an absorber length of 50 μm, and 2separate gain sections. The gain sections were biased togetherat 410 mA and the absorber was biased with −0.66 V.This resulted in Gaussian-shaped output pulses with 7.1picoseconds FWHM pulsewidth, 0.5 nm FWHM spectralwidth, and approximately 10 dB ER. The time bandwidthproduct was 0.43. The peak power was 6.8 dBm onchip,determined from onchip photocurrent measurements.

With a 30.4 GHz, 13 dBm RF signal applied to the SAsection, this laser had 364 femtoseconds of absolute jitterand 50 MHz locking range. The laser could also be synchro-nized to 30 Gbps optical input signals with average powersbelow 0 dBm in the input waveguide meaning that it canperform all-optical clock recovery. To test this application,we intentionally degraded a 30.4 Gbps return to zero 231-1 pseudorandom bit stream by adding timing jitter andreducing the extinction ratio, and we used the racetrack ML-SEL to recover the clock all-optically [31]. The input ER was3.8 dB and the input jitter was 14 picoseconds. The recoveredclock had an ER of 10.4 dB and jitter of 1.7 picoseconds.The input data and output clock eye diagrams are shown inFigure 16(b). The device’s extreme regenerative capabilitiesand potential for integration with other components indicatethat it could be used as part of a silicon integrated optical 3Rregenerator.

Page 12: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

12 Advances in Optical Technologies

Gain 1 P contact Gain 1 N contact

SA

Gain 2P contact

Gain 2N contact

Anti-reflectionabsorber

Directional coupler SOA

1.45 mm

Data in

Clock out

Degraded 30.4 Gbps data in

Retimed, reshaped clock out

Figure 16: (a) SEM image of the 30.4 GHz racetrack ML-SEL. (b) Eye diagrams of the input data sent to the device and the regenerated,all-optically recovered clock.

Silicon waveguide n-InP layer

Input 1

TaperedIII-V junction

Input 2

Hybrid silicon evanescent amplifier Hybrid siliconevanescent photodetector

P pad N padN pad

P pad

(a)

500 μm

(b)

(c)

(d)

20 μm

5 μm

Figure 17: (a) Top view of the integrated device. (b) SEM image of eight fabricated devices. (c) Close view of the III-V taper of the amplifier.(d) Close view of the narrow III-V mesa of the detector.

4.2. Silicon evanescent amplifiers and photodetectors

Optical amplifiers are key components in realizing highlevels of photonic integration as they compensate for opticallosses from individual photonic elements. The hybrid sil-icon evanescent amplifier structure is similar to the offsetquantum well structure which has typical quantum wellconfinement factor in the range of 2% to 4% and it is suitablefor preamplifiers [35]. In this section, the integration of anamplifier and a detector for improved receiver sensitivity isdiscussed [36].

Figure 17(a) shows a device structure of the integrateddevice with an amplifier and a detector. At the transitionbetween the passive silicon waveguide and the hybridwaveguide of the amplifier, the width of the III-V mesa istapered from 0 μm to 4 μm over a length of 70 μm to increasethe coupling efficiency and to minimize reflection. The widthfrom 4 μm to 14 μm is tapered more abruptly over 5 μm

since III-V mesas wider than 4 μm do not laterally affect theoptical mode. A 7◦ tilted abrupt junction is used between thepassive silicon waveguide and the detector hybrid waveguide.The details of the epitaxial structure of the III-V layersare summarized in Table 1. The III-V mesa width of theamplifier is 14 μm. The III-V mesa width of the detector is3 μm at the p cladding layer and 2 μm at the p SCH and thequantum well layers to reduce the capacitance of the device.The detector p and n pads are designed to be 100 μm apartfrom center to center to use a standard GSG RF probe forhigh-speed testing. The SEM image of the eight fabricateddevices is shown in Figure 17(b), and the close view of theIII-V amplifier taper and the III-V detector mesa are shownin Figures 17(c) and 17(d) respectively. The total lengthof the amplifier and the detector is 1.24 mm and 100 μm,respectively.

The internal quantum efficiency is measured to bearound 50% at a reverse bias voltage of 2 V at 1550 nm as

Page 13: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 13

0.65

0.6

0.55

0.5

0.45

0.4

0.35

0.3

0.250 1 2 3 4

Reverse bias (V)

Inte

rnal

quan

tum

effici

ency

1550 nm1557 nm

(a)

10

5

0

−5

−10

−15

−20100 200 300 400

Current (mA)

SOA

chip

gain

(dB

)

1550 nm1557 nm

(b)

Figure 18: (a) Internal quantum efficiency of the detector. (b) Gain of the amplifier.

30

25

20

15

10

5

00 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Coupled input power (mW)

Det

ecto

rcu

rren

t(m

A)

R = 5.7 A/W

0.5 dB

Figure 19: Saturation characteristics of the integrated opticallypreamplified receiver.

shown in Figure 18(a). The amplifier gain is also measuredby taking photocurrent from the detector and the maximumgain is 9.5 dB at 300 mA as shown in Figure 18(b). Thereflection from the III-V taper is estimated from the ripplesat the ASE spectrum and it is less than 5 × 10−4. Thetaper loss is estimated to be in the range of 0.6 dB to1.2 dB by measuring the photocurrent from the reverse-biased amplifier [36]. Figure 19 shows the saturation char-acteristics of the integrated device. The overall responsivityis 5.7 A/W and the device is saturated by 0.5 dB at an outputphotocurrent of 25 mA. The device bandwidth is measuredto be 3 GHz from the time domain impulse measurementseven though the estimated RC limited bandwidth is 7.5 GHz.

10−4

10−5

10−6

10−7

10−8

10−9

10−10

−22 −20 −18 −16 −14 −12 −10 −8

Coupled input power (dBm)

BE

R@

2.5

Gbp

sN

RZ

PR

BS

300 mA200 mA150 mA

100 mANo Amp

8.5 dB

Figure 20: Optically preamplified receiver bit error rate curvesmeasured with 231-1 NRZ 2.5 Gbps transmissions with differentamplifier gains.

The current bandwidth is limited by carrier trapping inthe quantum wells. Higher bandwidth can be achieved byincorporating a thinner SCH layer and a bulk absorbing

Page 14: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

14 Advances in Optical Technologies

100μm

Figure 21: Photograph of the fabricated EAM modulator.

2

0

−2

−4

−6

−8

−10

−12

−14

0 1 2 3 4 5 6

Reverse bias (V)

Rel

ativ

eex

tin

ctio

n(d

B)

Figure 22: EAM extinction versus bias.

region or new quantum wells with small valence band offset.The bit error rate (BER) was measured with 2.5 Gbps NRZ231-1 pseudorandom bit sequence (PRBS) with differentamplifier gains and the result is shown in Figure 20. Thepurple data points are baseline measurements without theamplification by launching the signal to input 2. The BERdata at an amplifier current of 100 mA shows worse receiversensitivity than the receiver sensitivity without amplification(baseline) because the amplifier is below transparency. Oncethe amplifier is driven beyond transparency, the powerpenalty becomes negative as shown in the three BER curveson the left side. At the maximum gain of 9.5 dB, the powerpenalty is −8.5 dB compared to the baseline and the receiversensitivity at a BER of 10−9 is−17.5 dBm. The 1 dB differencebetween the gain and the measured power penalty is dueto the ASE noise. Better sensitivities would be achievablewith a good transimpedance amplifier and this device canbe integrated with silicon passive wavelength demultiplexersfor high-speed WDM receivers [37].

4.3. Silicon evanescent electroabsorption modulators

In this section, we review hybrid silicon evanescent elec-troabsorption modulators (EAMs). The modulator structuredescribed here can be integrated with lasers, amplifiers,and photodetectors using quantum well intermixing [38]

2

0

−2

−4

−6

−8

−100 5 10 15 20

Frequency (GHz)

Nor

mal

ized

RF

pow

er(d

B)

Figure 23: Measured small signal response (solid). Calculatedresponse using an RC model (dashed line).

enabling integrated high-speed transmitters. The cross-sectional structure is shown in Figure 1(b) and the III-Vepitaxial structure with photoluminescence at 1478 nm issummarized in Table 3. AlGaInAs is chosen as the multiplequantum well material because typically it has a largeconduction band offset which provides a stronger carrierconfinement and produces strong quantum confined Starkeffect with higher extinction ratio [39, 40]. The siliconwaveguide was fabricated with a height of 0.5 μm and aslab thickness of 0.3 μm. The silicon waveguide has a widthof 1.5 μm for passive segments and is tapered to 0.8 μmin the hybrid modulator region for a larger quantum wellconfinement factor. The width of the III-V mesa is 4 μm atthe top InP cladding layer and 2 μm at the SCH and quantumwell layers to reduce the capacitance of the device [40, 41].The overall layout of the device is the same as the detectordescribe in the previous section except for the width of theIII-V mesa is tapered from 0 to 2 μm over a length of 60 μmto increase the coupling efficiency and to minimize reflectionas in the amplifiers. The hybrid EAM has a total lengtharound 220 μm with 100 μm absorber and two 60 μm longtapers. The photograph of the fabricated device is shown inFigure 21.

Figure 22 shows the relative extinction at wavelengthof 1550 nm under various reverse biases. More than 10 dBextinction can be achieved with a reverse bias voltage of 5 V.The device has a series resistance around 30Ω and capaci-tance of 0.1 pF measured from the impedance measurementsresulting in a RC limited bandwidth of ∼20 GHz. It matcheswith the measured small signal modulation response asshown in Figure 22.

To investigate the performance of large signal modula-tion, the modulator is driven with a 231-1 pseudorandombit sequence (PRBS). The device is biased at −3 V with apeak-to-peak drive voltage of 3.2 V. The modulated light iscollected with a lensed fiber and amplified with an EDFA.Figure 24(a) shows eye diagrams measured at nonreturn-to-zero (NRZ) 10 Gbps. The 10 Gbps signal has an extinction

Page 15: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 15

Table 3: III-V epitaxial layer structure with a 1478 nm photoluminescence peak.

Name Composition Doping concentration Thickness

P contact layer P-type In0.53Ga0.47As 1× 1019 cm−3 0.1 μm

Cladding P-type InP 1× 1018 cm−3 1.5 μm

SCH P-type Al0.160Ga0.320In0.520As, 1.3Q 1× 1017 cm−3 0.15 μm

Quantum wellsAl0200Ga0.330In0.470As, 1.19Q(11x) Undoped 7 nm

Al0.080Ga0.330In0.590As, 1.55Q(10x) Undoped 11 nm

SCH Al0.160Ga0.320In0.520As, 1.3Q Undoped 0.1 μm

N layer N-type InP 1× 1018 cm−3 110 nm

Super latticeN-type In0.85Ga0.15As0.327P0.673, 1.1Q(2x) 1× 1018 cm−3 7.5 nm

N-type InP (2x) 1× 1018 cm−3 7.5 nm

N bonding layer N-type InP 1× 1018 cm−3 10 nm

26.5578 ns

Q-factor(cg)V base(cg)V top(cg)V amptd(cg)

Current11.66192.93 μW861.46 μW668.54 μW

Mean11.96198.73 μW854.22 μW655.50 μW

std dev1.163.80 μW8.69 μW10.47 μW

50 ps/div.

(a)

5

4

3

2

1

0

−1

−2

−30 100 200 300 400 500

Time (ps)

Am

plit

ude

(b)

Figure 24: (a) Measured 10G NRZ eye diagrams; (b) pulse trainof drive signal (dashed) and optical modulated signal (solid).The drive signal has rise/fall time around 35 picosesonds. Themodulated signal shows rising/falling times of about 27 picseconds.

ratio of 6.3 dB, which is slightly lower than the DC extinctiondue to additional microwave voltage drop at cladding andohmic contacts. The eye is clearly open with quality factor(Q factor) close to 12. The rise and fall times of the signal are

about 27 picoseconds, which is, as expected, faster than thedriving signal as shown in Figure 24(b).

5. CONCLUSION

Recent progress of hybrid silicon evanescent devices has beenreviewed in this paper. Discrete lasers, amplifiers, photode-tectors, and electroabsorption modulators have been demon-strated. Racetrack lasers integrated with photodetectors,mode locked lasers, and photodetectors with preamplifiershave also been presented as examples of photonic integrationwith this hybrid device structure. These demonstrationsshow the potential for realizing active functionality on thesilicon photonics platform. One of the important pathsof this research is improving performance of individualdevices, that is, device efficiency and thermal performance,in conjunction with studies on device reliability. Anotherpath is the development of bonding of III-V materials tolarge size silicon wafers (>6 inch). The bonding processcan be wafer scale or can be used for die attach. Theoptimum size of III-V material to use depends on the densityof active devices required for the silicon wafer [42]. Thehybrid silicon evanescent device platform provides a uniqueway to build photonic active devices on silicon, and thosestudies will expedite the applications of silicon photonicintegrated circuits in optical telecommunications and opticalinterconnects.

ACKNOWLEDGMENTS

The authors would like to thank M. J. Paniccia, J. Shah,M. Haney, and W. Chang for insightful discussions andO. Cohen and O. Raday for silicon fabrication. This workwas supported by DARPA/MTO and ARL under AwardsW911NF-05-1-0175 and W911NF-04-9-0001 and by IntelCorporation.

REFERENCES

[1] F. Xia, M. Rooks, L. Sekaric, and Y. Vlasov, “Ultra-compacthigh order ring resonator filters using submicron siliconphotonic wires for on-chip optical interconnects,” OpticsExpress, vol. 15, no. 19, pp. 11934–11941, 2007.

Page 16: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

16 Advances in Optical Technologies

[2] F. Xia, L. Sekaric, and Y. Vlasov, “Ultracompact optical bufferson a silicon chip,” Nature Photonics, vol. 1, no. 1, pp. 65–71,2007.

[3] M. Loncar, T. Doll, J. Vuckovic, and A. Scherer, “Design andfabrication of silicon photonic crystal optical waveguides,”Journal of Lightwave Technology, vol. 18, no. 10, pp. 1402–1411, 2000.

[4] P. D. Trinh, S. Yegnanarayanan, F. Coppinger, and B.Jalali, “Silicon-on-insulator (SOI) phased-array wavelengthmulti/demultiplexer with extremely low-polarization sensitiv-ity,” IEEE Photonics Technology Letters, vol. 9, no. 7, pp. 940–942, 1997.

[5] K. Jia, W. Wang, Y. Tang, et al., “Silicon-on-insulator-basedoptical demultiplexer employing turning-mirror-integratedarrayed-waveguide grating,” IEEE Photonics Technology Let-ters, vol. 17, no. 2, pp. 378–380, 2005.

[6] A. Liu, R. Jones, L. Liao, et al., “A high-speed silicon opticalmodulator based on a metal-oxide-semiconductor capacitor,”Nature, vol. 427, no. 6975, pp. 615–618, 2004.

[7] D. Marris-Morini, X. L. Roux, L. Vivien, et al., “Opticalmodulation by carrier depletion in a silicon PIN diode,” OpticsExpress, vol. 14, no. 22, pp. 10838–10843, 2006.

[8] Y. Jiang, W. Jiang, L. Gu, X. Chen, and R. T. Chen, “80-micron interaction length silicon photonic crystal waveguidemodulator,” Applied Physics Letters, vol. 87, no. 22, Article ID221105, 3 pages, 2005.

[9] Q. Xu, B. Schmidt, S. Pradhan, and M. Lipson, “Micrometre-scale silicon electro-optic modulator,” Nature, vol. 435, no.7040, pp. 325–327, 2005.

[10] R. S. Jacobsen, K. N. Andersen, P. I. Borel, et al., “Strainedsilicon as a new electro-optic material,” Nature, vol. 441, no.7090, pp. 199–202, 2006.

[11] J. E. Roth, O. Fidaner, R. K. Schaevitz, et al., “Opticalmodulator on silicon employing germanium quantum wells,”Optics Express, vol. 15, no. 9, pp. 5851–5859, 2007.

[12] D. Ahn, C.-Y. Hong, J. Liu, et al., “High performance,waveguide integrated Ge photodetectors,” Optics Express, vol.15, no. 7, pp. 3916–3921, 2007.

[13] A. O. Splett, T. Zinke, B. Schueppert, et al., “Integratedoptoelectronic waveguide detectors in SiGe for optical com-munications,” in Photodetectors and Power Meters II, vol. 2550of Proceedings of SPIE, pp. 224–234, San Diego, Calif, USA,July 1995.

[14] G. Masini, G. Capellini, J. Witzens, and C. Gunn, “A four-channel, 10 Gbps monolithic optical receiver in 130 nmCMOS with integrated Ge waveguide photodetectors,” inProceedings of the Optical Fiber Communication and theNational Fiber Optic Engineers Conference (OFC/NFOEC ’07),Anaheim, Calif, USA, March 2007, PDP 31.

[15] R. Jones, H. Rong, A. Liu, et al., “Net continuous wave opticalgain in a low loss silicon-on-insulator waveguide by stimulatedRaman scattering,” Optics Express, vol. 13, no. 2, pp. 519–525,2005.

[16] H. Rong, R. Jones, A. Liu, et al., “A continuous-wave Ramansilicon laser,” Nature, vol. 433, no. 7027, pp. 725–728, 2005.

[17] O. Boyraz and B. Jalali, “Demonstration of a silicon Ramanlaser,” Optics Express, vol. 12, no. 21, pp. 5269–5273, 2004.

[18] S. G. Cloutier, P. A. Kossyrev, and J. Xu, “Optical gain andstimulated emission in periodic nanopatterned crystallinesilicon,” Nature Materials, vol. 4, no. 12, pp. 887–891, 2005.

[19] Z. Mi, P. Bhattacharya, J. Yang, and K. P. Pipe, “Room-temperature self-organised In0.5Ga0.5As quantum dot laser onsilicon,” Electronics Letters, vol. 41, no. 13, pp. 742–744, 2005.

[20] D. Pasquariello and K. Hjort, “Plasma-assisted InP-to-Si lowtemperature wafer bonding,” IEEE Journal on Selected Topicsin Quantum Electronics, vol. 8, no. 1, pp. 118–131, 2002.

[21] Y. C. Liu, L. T. Ho, Y. B. Bai, et al., “Growth of ultrathin SiO2

on Si by surface irradiation with an O2+ Ar electron cyclotronresonance microwave plasma at low temperatures,” Journal ofApplied Physics, vol. 85, no. 3, pp. 1911–1915, 1999.

[22] U. Gosele and Q.-Y. Tong, “Semiconductor wafer bonding,”Annual Review of Materials Science, vol. 28, no. 1, pp. 215–241,1998.

[23] A. W. Fang, H. Park, O. Cohen, R. Jones, M. J. Paniccia, andJ. E. Bowers, “Electrically pumped hybrid AlGaInAs-siliconevanescent laser,” Optics Express, vol. 14, no. 20, pp. 9203–9210, 2006.

[24] L. Coldren and S. Corzine, Diode Lasers and Photonic Inte-grated Circuits, John Wiley & Sons, New York, NY, USA, 1995.

[25] M. N. Sysak, H. Park, A. W. Fang, et al., “Experimental andtheoretical thermal analysis of a hybrid silicon evanescentlaser,” Optics Express, vol. 15, no. 23, pp. 15041–15046, 2007.

[26] H.-H. Chang, A. W. Fang, M. N. Sysak, et al., “1310 nm siliconevanescent laser,” Optics Express, vol. 15, no. 18, pp. 11466–11471, 2007.

[27] Y.-A. Chang, T.-S. Ko, J.-R. Chen, et al., “The carrier blockingeffect on 850 nm InAlGaAs/AlGaAs vertical-cavity surface-emitting lasers,” Semiconductor Science and Technology, vol. 21,no. 10, pp. 1488–1494, 2006.

[28] A. W. Fang, R. Jones, H. Park, et al., “Integrated AlGaInAs-silicon evanescent race track laser and photodetector,” OpticsExpress, vol. 15, no. 5, pp. 2315–2322, 2007.

[29] H. Park, A. W. Fang, R. Jones, et al., “A hybrid AlGaInAs-silicon evanescent waveguide photodetector,” Optics Express,vol. 15, no. 10, pp. 6044–6052, 2007.

[30] B. R. Koch, A. W. Fang, O. Cohen, and J. E. Bowers, “Mode-locked silicon evanescent lasers,” Optics Express, vol. 15, no. 18,pp. 11225–11233, 2007.

[31] B. R. Koch, A. W. Fang, H. N. Poulsen, et al., “All-opticalclock recovery with retiming and reshaping using a siliconevanescent mode locked ring laser,” in Proceedings of theOptical Fiber Communication and the National Fiber OpticEngineers Conference (OFC/NFOEC ’08), San Diego, Calif,USA, February 2008.

[32] E. A. Avrutin, J. H. Marsh, and E. L. Portnoi, “Monolithic andmulti-gigahertz mode-locked semiconductor lasers: construc-tions, experiments, models and applications,” IEE Proceedings:Optoelectronics, vol. 147, no. 4, pp. 251–278, 2000.

[33] P. J. Delfyett, S. Gee, M.-T. Choi, et al., “Optical frequencycombs from semiconductor lasers and applications in ultra-wideband signal processing and communications,” Journal ofLightwave Technology, vol. 24, no. 7, pp. 2701–2719, 2006.

[34] M. Teshima, K. Sato, and M. Koga, “Experimental investiga-tion of injection locking of fundamental and subharmonicfrequency-modulated active mode-locked laser diodes,” IEEEJournal of Quantum Electronics, vol. 34, no. 9, pp. 1588–1596,1998.

[35] H. Park, A. W. Fang, O. Cohen, R. Jones, M. J. Paniccia, and J.E. Bowers, “A hybrid AlGaInAs-silicon evanescent amplifier,”IEEE Photonics Technology Letters, vol. 19, no. 4, pp. 230–232,2007.

[36] H. Park, Y.-H. Kuo, A. W. Fang, et al., “A hybrid AlGaInAs-silicon evanescent preamplifier and photodetector,” OpticsExpress, vol. 15, no. 21, pp. 13539–13546, 2007.

Page 17: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

Hyundai Park et al. 17

[37] R. Nagarajan, M. Kato, S. Hurtt, et al., “Monolithic, 10and 40 channel InP receiver photonic integrated circuitswith on-chip amplification,” in Proceedings of the OpticalFiber Communication and the National Fiber Optic EngineersConference (OFC/NFOEC ’07), Anaheim, Calif, USA, March2007, PDP 32.

[38] J. W. Raring, E. J. Skogen, L. A. Johansson, et al., “Demonstra-tion of widely tunable single-chip 10-Gbps laser-modulatorsusing multiple-bandgap InGaAsP quantum-well intermixing,”IEEE Photonics Technology Letters, vol. 16, no. 7, pp. 1613–1615, 2004.

[39] J. Shimizu, M. Aoki, T. Tsuchiya, et al., “Advantages of opticalmodulators with InGaAlAs/InGaAlAs MQW structure,” Elec-tronics Letters, vol. 38, no. 15, pp. 821–822, 2002.

[40] H. Fukano, T. Yamanaka, M. Tamura, and Y. Kondo, “Very-low-driving-voltage electroabsorption modulators operatingat 40 Gbps,” Journal of Lightwave Technology, vol. 24, no. 5,pp. 2219–2224, 2006.

[41] K. Kato, A. Kozen, Y. Muramoto, Y. Itaya, T. Nagatsuma,and M. Yaita, “110-GHz, 50%-efficiency mushroom-mesawaveguide p-i-n photodiode for a 1.55-μm wavelength,” IEEEPhotonics Technology Letters, vol. 6, no. 6, pp. 719–721, 1994.

[42] J. E. Bowers, A. W. Fang, H. Park, R. Jones, M. J. Paniccia,and O. Cohen, “Hybrid silicon evanescent laser in a silicon-on-insulator waveguide,” in Proceedings of the Optical FiberCommunication and the National Fiber Optic Engineers Confer-ence (OFC/NFOEC ’07), pp. 1–3, Anaheim, Calif, USA, March2007, OTuK4.

Page 18: Photonic Integration on the Hybrid Silicon Evanescent ... · on hybrid silicon evanescent devices. In Section 2, the device structure and design issues are introduced. In Section

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttp://www.hindawi.com Volume 2010

RoboticsJournal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Active and Passive Electronic Components

Control Scienceand Engineering

Journal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

International Journal of

RotatingMachinery

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporation http://www.hindawi.com

Journal ofEngineeringVolume 2014

Submit your manuscripts athttp://www.hindawi.com

VLSI Design

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Shock and Vibration

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Civil EngineeringAdvances in

Acoustics and VibrationAdvances in

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawi Publishing Corporation http://www.hindawi.com

Volume 2014

The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014

SensorsJournal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Modelling & Simulation in EngineeringHindawi Publishing Corporation http://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Chemical EngineeringInternational Journal of Antennas and

Propagation

International Journal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Navigation and Observation

International Journal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

DistributedSensor Networks

International Journal of